The hypothesis that lonely children show hypervigilance for social threat was examined in a series of three studies that employed different methods including advanced eye-tracking technology. Hypervigilance for social threat was operationalized as hostility to ambiguously motivated social exclusion in a variation of the hostile attribution paradigm (Study 1), scores on the Children's Rejection-Sensitivity Questionnaire (Study 2), and visual attention to socially rejecting stimuli (Study 3). The participants were 185 children (11 years-7 months to 12 years-6 months), 248 children (9 years-4 months to 11 years-8 months) and 140 children (8 years-10 months to 12 years-10 months) in the three studies, respectively. Regression analyses showed that, with depressive symptoms covaried, there were quadratic relations between loneliness and these different measures of hypervigilance to social threat. As hypothesized, only children in the upper range of loneliness demonstrated elevated hostility to ambiguously motivated social exclusion, higher scores on the rejection sensitivity questionnaire, and disengagement difficulties when viewing socially rejecting stimuli. We found that very lonely children are hypersensitive to social threat.